Market Discipline and Financial Crisis Policy: An Historical Perspective *
Michael D. Bordo
Rutgers University and NBER
June 2003
Paper prepared for the Contemporary Economic Policy Session: Market Discipline in Banking: Theory, and Evidence. Western Economic Association International Meetings, Denver Colorado, July 13 2003.
* For extremely stimulating discussions on this topic I would like to thank Donald Mathieson of the International Monetary Fund. For helpful comments on a previous draft I thank Anna Schwartz.
Market Discipline and Financial Crisis Policy: An Historical Perspective
1. Introduction
This paper presents a framework to aid in our understanding of the historical evolution of economic policy to deal with financial crises. There are three pillars of financial crisis policy: crisis prevention, crisis management and crisis resolution.1 The framework applies most readily to domestic financial crisis policy but also has resonance for international crises.
We define crisis prevention as measures taken by both the private and the public sector to make the financial system robust to financial crises. It includes measures taken by the private banking sector to build adequate capital and reserves to provide the wherewithal to meet sudden demands for liquidation; institutional structures to allow adequate risk pooling and portfolio diversification such as nationwide branch banking systems; techniques to reduce the information asymmetry that makes it difficult to distinguish
sound from unsound institutions such as uniform accounting standards; and government regulations which serve the same purposes such as deposit insurance, reserve and capital requirements.
- For an earlier use of the pillars concept see Fischer (1999).
Were crisis prevention successful, there would be no financial crises. In historical experience this has not happened. Once crises are triggered, crisis management comes into play. We define it as the extension of Walter Bagehot's (1873) famous strictures to central banks to lend freely, to illiquid but solvent banks, but at a penalty rate. Crisis management, however, can also be performed by private sector agents and governmental authorities other than the central bank.
Crisis resolution involves resolving the problems of various types of financial crises. It includes the liquidation of insolvent financial institutions, fiscal restructuring, default debt resolution and devaluation to resolve a currency crisis.
A large crisis can trigger changes in one or more of the pillars. In our framework, however, an innovation in one of the three underlying pillars can induce institutional changes in the financial system that may imply that further changes in the other pillars may be needed if the financial system is to remain stable and resistant to shocks. If these changes are not made, the resulting inconsistencies can create incentives for the market participants that lead to actions that set the stage for future crises. Thus, we can conceive
of a pendulum swinging between the pillars in the aftermath of a crisis.
An example that we develop below of how the pendulum process works is of a financial crisis arising in a country, triggered by an external shock, because the preventive measures in place are inadequate. The experience of this crisis in turn could generate innovations in one of the pillars. For example, new preventive measures that could have unforeseen consequences for the other pillars, (such as the extensive financial regulations of the New Deal era in the U.S.) which sowed the seeds for a new form of crisis several decades later and led to innovations in both crisis management and crisis resolution.
In the rest of the paper we first develop the pendulum hypothesis distinguishing between domestic and international aspects. We then look at the historical record in some detail for the UK and U.S. and (briefly) other advanced countries. Then we focus on the emerging nations. We conclude with some lessons from the historical record.
- The Pendulum Hypothesis
2.1The Domestic Dimension
According to the pendulum hypothesis, crisis prevention, crisis management, and crisis resolution policies have evolved in response to the incentives they have created for market participants and to the occurrence of major shocks and/or crises that have signaled a weakening of one of the pillars.2
- Our concept of a pendulum is clearly related to the concept of a dialectic discussed by Kane (1991). The pendulum can also be thought as similar to Foucault’s pendulum, first displayed at the Pantheon in Paris in 1851 and currently on display at the Musee des arts et metiers. Foucault’s pendulum suspended from the ceiling was designed to demonstrate the rotation of the earth.
The pendulum swings when a crisis, triggered by the conjuncture of internal
inconsistencies in the financial regulatory system and external shocks reveals an
inconsistency between pillars. Innovation in one of the three underlying pillars creates incentives for market participants that lead to institutional changes that requires adjustments in the other pillars if financial policies are to effectively minimize and manage financial crises.
Thus in the history of the U.K. and the U.S. and other advanced countries, the evolution of the money economy and the use of fiduciary bank liabilities (whether convertible into specie or not) led to the problem of bank panics because of the inherent potential for instability in institutions the maturity of whose portfolios is mismatched. The potential
for losses to depositors/or noteholders and disruption of the payments system and hence
the real economy revealed a significant externality associated with banking. The externality was in part internalized by actions taken by the banks themselves (the use of reserve, capital and liquidity ratios) and collectively (the pooling of reserves) i.e. by crisis prevention,3 and later by the development of a central bank. The central bank, acting as a lender of last resort, became the crisis manager, either lending freely to illiquid but solvent financial institutions at a penalty rate on the basis of sound collateral or by the use
of open market operations to provide liquidity to the money market.
Typically a central bank evolved from a large government chartered and mandated bank, concerned primarily with its own commercial interest, to a central bank concerned with
the public interest (Goodhart 1988). Once a central bank began providing liquidity to the
financial system in the face of crisis, in a credible way, the externality of a banking panic was greatly reduced.4
Measures of crisis prevention have changed drastically over time: private clearing arrangements before 1914; to extensive interest rate ceilings and credit controls from the 1930’s through the 1970’s; to market-oriented capital requirements today. Crisis
management involving the central bank is no longer limited to Bagehot’s rule. The “too big to fail” doctrine and the concept of creative ambiguity have become prominent. Crisis resolution has evolved from informally arranged mergers and lifeboat operations to deposit insurance and resolution agencies.
Taking the U.S. experience as a prototype, the first swing of the pendulum occurred in 1914 from earlier crisis prevention by both the private sector and the states and management by private arrangements and the U.S. Treasury in the face of a banking panic, to crisis management by a central bank - - the Federal Reserve.
- The practice of the private clearinghouse in the U.S. of pooling the reserves of the members and precommitting on eligibility was both a crisis prevention and crisis management device.
- Debate swirls over whether a central bank could ever do this – that it is very difficult to distinguish between solvency and illiquidity (Rockoff 1986) and that denying support to a large institutional bank can cause more harm to the real economy (Goodhart 1985).
However, once the lender of last resort was perceived to be in place the private financial sector reacted to it by reducing its safeguards (private crisis prevention mechanisms): a reduction in capital and liquidity ratios and dismantling of clearinghouse pooling arrangements. The inadequacies of this system were revealed when the system was hit by the massive shock of the Great Depression and the lender of last resort failed to deal with a crisis affecting a less prepared banking system. This crisis led in successive decades to a swing in the pendulum in the form of extensive regulation of the lending and investment activities of the private financial sector and the introduction of deposit insurance. The crisis management technique also changed because the central bank did not have the resources to recapitalize the financial system. Formal crisis resolution in the form of the Reconstruction Finance Corporation developed.
The next swing in the pendulum involved a dramatic change in crisis prevention.5 The end of interest rate ceilings and other domestic regulations but the survival of the safety net (including deposit insurance and the unwillingness to close large insolvent banks), exposed the financial system to moral hazard, increased risk taking and new forms of financial crisis.
The last swing in the pendulum has occurred in reaction to the moral hazard engendered by the safety net and the crisis resolution mechanism, in the direction of a restoration of the private techniques of providing a cushion of capital used before the advent of formal crisis management.
2.2.The International Dimension
The pendulum hypothesis applies to the international economy just as it does to the domestic economy. The three pillars of crisis policy, however, evolved differently. Crisis management evolved from the domestic lender of last resort function under the international gold (specie) standard.
Under a pure gold standard the monetary authority needs to supply gold coins to satisfy an unusual demand for currency. These can come from previously acquired international reserves or can be borrowed temporarily from another monetary authority, as was the case before 1914 for a number of rescues arranged between the Bank of England and the Banque de France (Bordo and Schwartz, 1998). In a mixed currency system with both gold coin and convertible fiduciary money, as was the case under the classical gold standard, crisis management applied the innovation of Bagehot’s Rule to lend freely on the basis of sound collateral but at a penalty rate. The monetary authority needed the credibility to temporarily suspend gold convertibility (in the UK invoke a Treasury letter) and to provide unlimited liquidity to the money market; otherwise gold had to be borrowed (recycled) from abroad.
5. It occurred in the 1960’s, 70’s and 80’s in the face of several historical forces: inflation; the growth of real income; reduced information asymmetries; the growth of technology which has improved access to information and led to the development of new types of financial assets; and arbitrage to exploit wedges the open up between free market returns and those suppressed by controls. These forces interacted with an ongoing tension that developed between the desire for stability which led to the creation of a financial safety net and market forces which have undermined the controls which buttress it.
Countries lacking such credibility had to abandon gold convertibility. Hence the international dimension of the lender of last resort function (a partial international lender of last resort) essentially involved the recycling of gold between countries.
There has never been an international lender of last resort facility capable of supplying high-powered money to the international monetary system in the way that a national lender of last resort in a fiat money regime can (Capie 1998). Only a supernational central bank as Keynes (1931) once advocated could serve as an international lender of last resort. Few countries, with the exception of the European Monetary Union, have expressed willingness to give up the monetary sovereignty required to a supernational central bank.
However, there have been periodic attempts at crisis management in which hard currencies were advanced to monetary authorities facing speculative attacks on their international reserves. These range from ad hoc intercentral bank lending and policy coordination under the gold standard and the interwar gold exchange standard, to the rescues by the G-10 and the IMF during the Bretton Woods era, to the recent massive infusions of liquidity by the IMF and major countries to emerging countries in crisis today. These rescues have not closely followed Bagehot’s strictures. However, IMF conditionally has been viewed as an attempt to do so (Fischer 2000).
Crisis prevention in the international context in the pre 1914 period involved arrangements between the international branches and head office of commercial banks and gold policy by central banks to widen the gold points and thereby temporarily shield the domestic economy from external shocks that could lead to crisis. New powerful international preventive techniques that evolved after World War I included a panoply of capital and exchange controls and other restrictions on current account convertibility. The objective of the measures was to seal off the domestic economy from external shocks.
International crisis resolution techniques before 1914 included the recapitalization of a major investment bank, Barings in 1890, backstopped by a British government guarantee, the cutting off of international lending in the event of sovereign debt defaults and workouts by private foreign bond holders councils. Government involvement was
minimal (Lindert and Morton 1989). In recent years international crises have been resolved by national guarantees of the large international money center banks in the 1982 Debt Crisis. And by international bailouts of monetary authorities in crisis which often cover the interest and principal of the loans (Bordo and Schwartz, 1999b). Debt crises, after the defaults of the 1930’s, have been resolved by government mediation which often involved official guarantees and which have greatly lengthened the process of resolution (Bordo and Schwartz 1999b).
The pendulum would swing between these international crisis policy pillars, as was the case in the domestic context. Under the classical gold standard system, currency crises reflecting temporary inconsistencies between internal conditions and the external peg, were managed by domestic central banks raising their discount rates and running down their international reserves or temporarily borrowing gold from other central banks. (Bordo and Schwartz, 1999a). This practice began to erode even before 1914 as nations began violating “the rules of the game” to protect their national economies from external shocks, and then broke down with World War I which ended gold convertibility and imposed exchange and capital controls (Obstfield and Taylor 1998).
After World War I the restoration of the prewar system in the 1920’s was short lived. In the face of the massive shocks of the Great Depression, the pendulum swung in favor of extensive crisis prevention in the form of universal controls on capital movements and controls on the free convertibility of currencies. The postwar Bretton Woods System
represented another swing of the pendulum. It reestablished current account convertibility and the IMF, set up in part to provide temporary assistance to member countries facing temporary current account imbalances, became a partial lender of last resort. International crisis management was strengthened in the 1960’s by the rescue packages arranged by the G-10. In these years the presence of capital controls and regulation of domestic financial markets led to evasion and innovation to circumvent the controls. Their removal and the opening up of international capital markets in turn led to a series of crises in the 1980’s and 1990’s which led to the most recent swing in the pendulum toward new forms of crisis resolution (Fischer 1999).
3. The Three Pillars of Crisis Policy: The Historical Evolution of Advanced Countries.
We survey the evolution within our pendulum framework over the past century of crisis prevention, crisis management and crisis resolution in advanced countries. We distinguish between the national and international levels.
3.1. 1870-1914
1
The origins of financial crisis management and the lender of last resort can be traced back to the Bank of England’s response to banking panics in the eighteenth and nineteenth
centuries. Until the last third of the century, the Bank of England would react defensively to a decline in its gold reserves, whether from a shock to the balance of payments or a domestic banking crisis, by raising its discount rate. The Bank, which was a private institution established in 1694 as the government’s fiscal agent, would try to protect its shareholders during financial crises. On many occasions by delaying or refusing to provide liquidity support to stressed financial institutions, the Bank’s actions worsened the crisis. On several occasions between 1825 and 1866, the Bank would supplement its gold reserves by temporarily borrowing gold from the Banque de France. On a few occasions, this action prevented the Bank from suspending gold convertibility. On no occasion was the rescue sufficient to prevent a banking panic (Bordo and Schwartz, 1999b; Flandreau, 1998).
After the crisis of 1825, the Bank was under increasing pressure from Parliament and from critics such as Walter Bagehot to accept responsibility for the liquidity of the